NuClide Research — exposed AI and LLM infrastructure
Survey totals at a glance
{{ hCats }}
CATEGORIES
{{ hPlatforms }}
PLATFORMS SURVEYED
{{ hExposures }}
CONFIRMED EXPOSURES
{{ hReports }}
FIELD REPORTS
ENTER THE ATLAS
Catalog
→33 platform categories, every platform tagged with the auth posture it ships with.
Findings
→Field reports, defender advisories and methodology insights, searchable by category.
Wires
→Dispatches and technical essays, the reasoning behind the findings, not survey output.
THE WIDER ECOSYSTEM
Incident registries and threat frameworks map AI risk in the abstract. NuClide is the ground truth they describe, the exposed infrastructure itself. Complementary views of the same problem.
Catalog — platform categories by exposure posture
⌕
Tier filter shows categories containing at least one platform of that posture. A open by design · A* auth-off default · B setup-gate · C hardened.
{{ c.n }}
{{ c.badge }}
{{ c.name }}
{{ c.note }}
{{ c.metricLabel }}
{{ c.rateText }}
Field reports — findings, advisories and methodology
⌕
{{ postCountText }}
EVIDENCE
Evidence rung of a finding. A direct 200-with-data read · A* corroborated read · B indirect or single-signal. ALL includes the {{ ctTierUntiered }} reports not yet rung-graded.